Michau/t5-base-en-generate-headline
Full fine-tune vs google-t5/t5-base. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Attention 147/147, Norms 62/62, other 48/48
Full fine-tune vs google-t5/t5-base. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Attention 147/147, Norms 62/62, other 48/48